game variants
NLHE and PLO4 use independent validated showdown paths.
Independent poker research software
Poker study,
measured.
Study strategy, run transparent solves, train decisions, analyze your own hands, and play continuous games against configurable AI—all in one disciplined workspace.
Educational study only. PokerLab never facilitates real-money wagering.
Solution explorerNLHE · BTN vs BB
K♠7♥2♣
5.5 BB pot
Selected combo
A♠ K♠
StrategyPending
Action EV—
Equity—
Numbers remain unavailable until a versioned solver artifact is attached.
Calculation provenance enforcedReference solver idle
canonical flops
Suit-isomorphic classes map back to all 22,100 physical flops.
AI fidelity
Policy divergence and estimated regret—not arbitrary random play.
One research system
Every mode speaks the same mathematical language.
Strategy artifacts, hand states, ranges, actions, and analysis records move through documented, versioned boundaries.
Study
Navigate study trees from preflop through river with combo-level range inspection and explicit artifact provenance.
Explore module →Solve
Configure ranges, stacks, rake, bet trees, and nodelocks through a versioned asynchronous solver contract.
Explore module →Practice
Save and share drills and grade bounded exact terminal decisions; general node-based artifact feedback remains a future integration.
Explore module →Analyze
Parse owned hand histories, replay factual actions, and attach measured strategy comparisons without observing live clients.
Explore module →Play AI
Play continuous NLHE or four-card PLO sessions against calibrated, hidden-information-safe bot policies.
Explore module →Preflop
Inspect NLHE combos and PLO structure, save chart context, and keep unsupplied solved frequencies explicitly unavailable.
Explore module →Poker Math
Learn the mathematics underneath decisions through formulas, worked examples, calculators, and quizzes.
Explore module →Honest computation
No mystery numbers.
PokerLab marks a value as solved, approximated, sampled, or unavailable. Every result carries its algorithm, abstraction, iteration count, seed policy, and quality metric where applicable.
Open solver methodology →Deterministic rules
Seeded dealing, legal actions, pot-limit maxima, side pots, split pots, rake, ICM, and exact PLO card use are tested as pure state transitions.
Measured strategy
The built-in reference solver reports exact best-response exploitability for tractable Kuhn games. Larger poker remains an explicit external-compute dependency.
Calibrated opponents
AI fidelity mixes a versioned baseline with structured policy perturbations and reports divergence plus estimated regret.
Build better decisions
Bring the whole game into one lab.
Start with the limited working demo. Qualify monthly membership in Stripe test mode before any separately authorized live launch.